Novel approach: Naïve Bayes with Vector space model for spam classification

Abstract

We always see our normal mail goes into spam folder of the mail box. Interestingly 90% of the time the mail server classifies it perfectly but sometimes it fails due to spammer are getting highly technical. In this paper, we are using novel approach which uses Vector space model with Naïve Bayes to correctly classify mails as spam mail. Naïve Bayes method is used for spam classification but still binding with personalize word vector helps in increasing the accuracy of the system because user receives special type of message only. In this research work, we use vector space model with naïve bayes to classify spam mail. We got nearly 85% of accuracy in spam classification. We have used personalize mail classification option instead of standard global classification because people visiting subjective (i.e. pornographic) sites frequently get spam mail related to that subject (pornography) only and hence personalization shows improved result.